Mod_cluster does not follow manual ProxyPass strictly

Issue

• Mod_cluster does not follow manual ProxyPass strictly if using as follows with CreateBalancer set to 1 from single virtual host. For example.

NameVirtualHost *:8000
Listen 8000

 <VirtualHost *:8000>

  ProxyPass        /A/app1  balancer://lb1/A/app1 stickysession=JSESSIONID|jsessionid nofailover=On
  ProxyPassReverse /A/app1  balancer://lb1/A/app1

  ProxyPass        /A/app2  balancer://lb2/A/app2 stickysession=JSESSIONID|jsessionid nofailover=On
  ProxyPassReverse /A/app2  balancer://lb2/A/app2

 </VirtualHost>

• It is observed that modcluster does not honor this manual proxy pass. If all workers in lb1 are down, requests like http://localhost:8000/A/app1 are still going to JBoss node from lb2 balancer and giving HTTP 404 error. I expect in an event all workers in lb1 are down then request like http://localhost:8000/A/app1 should give HTTP 503 error.
• We have two balancers serving the same application contexts. We are using the ProxyPass directive to forward requests based on URIs to each of the balancers.
• However it is observed that when all members of a balancer are offline, then apache tries to route the request to the second balancer serving the same context root. This is happening despite the URI definition in the ProxyPass directive.
• Would this be a known issue or bug or the default apache proxy pass behaviour, as I could not find any documentation to indicate what happens when none of the proxypass patterns match.
• modcluster proxy routing error
• Routing requests to correct nodes with mod_cluster.